Journal Pre-proofs

Changes in aerobic fermentation and microbial community structure in foodwaste derived from different dietary regimes

Yanzeng Li, Zhou Chen, Yanyan Peng, Kaiming Zheng, Chengsong Ye, KunWan, Shenghua Zhang

PII: S0960-8524(20)31220-7DOI: https://doi.org/10.1016/j.biortech.2020.123948Reference: BITE 123948

To appear in: Bioresource Technology

Received Date: 31 May 2020Revised Date: 25 July 2020Accepted Date: 28 July 2020

Please cite this article as: Li, Y., Chen, Z., Peng, Y., Zheng, K., Ye, C., Wan, K., Zhang, S., Changes in aerobicfermentation and microbial community structure in food waste derived from different dietary regimes,Bioresource Technology (2020), doi: https://doi.org/10.1016/j.biortech.2020.123948

This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a coverpage and metadata, and formatting for readability, but it is not yet the definitive version of record. This versionwill undergo additional copyediting, typesetting and review before it is published in its final form, but we areproviding this version to give early visibility of the article. Please note that, during the production process, errorsmay be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

© 2020 Published by Elsevier Ltd.

Changes in aerobic fermentation and microbial community structure

in food waste derived from different dietary regimes

Yanzeng Li a, b, Zhou Chen a, b, Yanyan Peng a, Kaiming Zheng c, Chengsong Ye a, Kun

Wan a, b , Shenghua Zhang a, *

a Key Lab of Urban Environment and Health, Institute of Urban Environment, Chinese

Academy of Sciences, Xiamen 361021, China

b University of Chinese Academy of Science, Beijing 100049, China

c School of Astronautics, Harbin Institute of Technology, Harbin 150001, China

Abstract: This study aimed to analyze the relationship between food components and

food waste aerobic fermentation efficiency. Different food wastes were designed to be

reflective of different dietary regimes, including formulated (R1), high oil/fat and salt

(R2), high oil/fat and sugar (R3), and vegetarian (R4) diets, after which the

physicochemical properties, enzyme activity, and structural characteristics of food

waste microbial communities were examined to explore the potential mechanisms of

food waste degradation under different dietary regimes. The main results of this study

demonstrated that the physicochemical properties and hydrolase activity of different

food waste were significantly different. The species richness in R2 and R3 food waste

was higher than that of R1 and R4, whereas the community diversity of R1 and R4 food

waste was higher than that of R2 and R3. At the genus level, the dominant bacteria in

the four food waste types were Bacillus, Thermoactinomyces, Paenibacillus, and

Cohnella.

Keywords: food waste; aerobic fermentation; dietary regimes; physicochemical

properties; enzyme activity; microbial community

1. Introduction

Food waste (FW) is an important part of municipal solid waste (MSW) and

includes food residues and kitchen waste. FW is mainly derived from residues in

restaurants and cafeterias, as well as waste from food processing, including perishable

organic wastes such as leftovers, vegetables, and peels (Thyberg and Tonjes, 2016).

Relevant research indicates that approximately 1.3 billion tons of food waste are

generated globally each year and the amount of food waste is expected to increase in the

next 25 years (Chen et al., 2017; Petracchini et al., 2017). Food waste is a critical

component of environmental sustainability and can be a valuable resource if properly

processed (Conrad et al., 2018). Various treatment and disposal strategies have been

developed to utilize and recycle the current abundance of food waste. For example,

hydrogen, methane, lactic acid, and other substances have been obtained from food

waste anaerobic digestion (Li et al., 2019). Similarly, aerobic composting has been

shown to significantly decrease the weight and volume of food waste, which can

ultimately be transformed into highly fertile and stable humus (Bernal et al., 2009).

However, regardless of whether aerobic composting or wet or dry anaerobic

fermentation are used to process food waste, the treatment effectiveness is largely

dependent on the waste components. For instance, the outcomes of treating food waste

via dry anaerobic and wet anaerobic processes can vary depending on the solid contents

of the waste material. Li et al. (2019) reported differences in CH4 gas production during

the anaerobic digestion of different food waste components. Qureshi et al. (2017) also

reported differences in pectinase and lipase activity during the solid fermentation of

different fruit waste. Moreover, Wang et al. (2017) revealed differences in microbial

community structure and composting effects during aerobic composting, which were

attributed to differences in food waste feeding ratios. Different regions are affected by

varying climates and religious beliefs, thereby affecting the dietary habits of people. For

instance, the vast majority of Europeans and Americans follow so-called "western

diets," which are mainly characterized by higher intakes of sugar, salt, and fat (Conrad

et al., 2018). Moreover, people who follow this diet typically exhibit an increased risk

of certain diseases such as hypertension and diabetes, and it is estimated that 1.12

billion people worldwide will be obese by 2030 (Kastorini et al., 2011; Brigham et al.,

2015). In India, approximately 35% of the population follows a strictly vegetarian diet,

which has been passed down from generation to generation. Vegetarian diets are

generally thought to promote health. Compared with carnivores, vegetarians have lower

plasma cholesterol and lower risk of hypertension and obesity (Key et al., 2006).

However, regardless of eating habits (i.e., food composition), it is still unclear whether

generated food waste is easier or more difficult to process, and very few research has

focused on determining the effect of eating habits on food waste treatment and

management. Food waste composed of various components has its own unique

problems during treatment. High oil, salt, and fat food waste has been long known to

cause substantial problems during transportation and disposal, whereas vegetarian food

waste contains large amounts of cellulose and hemicellulose and therefore it cannot be

easily degraded during composting and fermentation. Therefore, it is important to

consider the eating habits of the population when designing food waste treatment

strategies.

The biochemical degradation of food waste is composed of a series of complex

biochemical reactions (Bernal et al., 2009). During the fermentation process,

environmental factors and microbial abundance, diversity, and community composition

of food waste are known to change significantly. Physicochemical properties are the

main factors influencing the degradation of food waste. Generally, as temperature rises

during the composting process, organic matter and moisture contents are significantly

reduced, whereas pH gradually increases with the decomposition of nitrogenous organic

matter such as protein (Bernal et al., 2009; Wang et al., 2019). A large number of

studies have assessed the composting process by monitoring changes in

physicochemical properties and hydrolytic enzyme activity. Chang and Hsu (2008)

sought to optimize composting conditions by exploring the dynamic changes in

physicochemical properties during the composting process. Awasthi et al. (2018)

studied the biological mechanisms of the composting process and found that hydrolytic

enzyme activity increased and then decreased during aerobic composting, which was

consistent with the findings of previous studies (Awasthi et al., 2017; Qureshi et al.,

2017).

Microbial agents play an important role in the aerobic fermentation of food waste.

Zhang et al. (2014) added lignocellulosic decomposers at the beginning of composting,

which greatly improved the waste degradation rate. Onwosi et al. (2017) reported that

the use of microbial agents can shorten the composting process and increase the density

of dominant bacteria. The biomass of indigenous microorganisms is small during the

early stages of composting, and thus its takes some time for these organisms to

reproduce (Xi et al., 2015). Therefore, it is necessary to add a microbial compound

agent to food waste prior to aerobic fermentation. Artificial addition of high-efficiency

microbial compound agents can adjust the structure of the microflora and improve

microbial activity, thus improving composting efficiency, shortening the fermentation

cycle, and improving the quality of composting products (Li et al., 2020). Bacterial

activity is closely related to the composting process, and the changes in the microbial

community structure of food waste can be determined via sequencing of bacterial 16S

rRNA genes. Numerous studies have demonstrated that the microbial community during

aerobic composting changes in stages with temperature (Wang et al., 2017; Nakasaki et

al., 2019). However, the effect of food waste composition associated with different

eating habits on microbial community structures remains unclear.

Therefore, this study sought to characterize the effect of food waste composition

derived from different eating habits on aerobic fermentation and microbial community

structure. In this study, food waste was generated to be representative of different eating

habits, including formulated (R1), high oil/fat and salt (R2), high oil/fat and sugar (R3),

and vegetarian (R4) diets to identify dynamic changes in bacterial community structure

and biochemical degradation mechanisms. The analysis and discussion of the

correlation of physicochemical indicators, hydrolytic enzyme activity, and microbial

community changes can directly reflect the characteristics of the aerobic fermentation of

food waste and the succession of microbial communities. Based on these results, the

biochemical degradation of food waste under four different dietary regimes was

characterized, with special attention to the ability of current fermentation strategies to

improve the efficiency of food waste aerobic fermentation.

2. Materials and Methods

2.1 Materials and experimental setup

The food waste was designed to be representative of formulated diet, high oil/fat

and salt, high oil/fat and sugar, and vegetarian diets, which were based on different

well-characterized dietary proportion structures (Food and Agriculture Organization of

the United Nations. 2016). Different ingredients were procured, cooked, and ground,

after which a microbial compound agent was added. This added EM solid microbial

agent (containing Bacillus, Actinomycetes, Yeast, Lactobacillus, and other

microorganisms) was obtained from China Agricultural University. The properties and

proportions of the feedstocks are summarized in Table 1.

The aerobic fermentation reactor was outfitted with a custom-designed thermostat

for peritoneal fluid with four cuboid iron boxes inside measuring 15 cm long × 8 cm

wide × 15 cm high (Fig. 1). In order to create a good aerobic state and facilitate airflow,

an aeration head (1.8 L h-1) was installed at the bottom of the reactor so that oxygen

could be introduced from bottom to top, which was combined with artificial stirring

twice a day at fixed time intervals. Afterward, the different food waste compositions

and microbial agents were fully mixed in the reactor at a 10:1 ratio to initiate aerobic

fermentation.

2.2 Sample collection

The waste heap was turned and sampled regularly to ensure that the food waste

remained in an aerobic fermentation state. Prior to the experiment, the food waste was

allowed to ferment in natural ventilation conditions, and each sample was made of a

mixture of parallel samples randomly collected from different cross-sections and depths

of each reactor on days 0, 8, 13, 18, 23 and 28. The samples were then divided into

three parts; one of the parts was dried in an oven at 30 ℃ and then used to determine the

physicochemical properties of food waste. Another part was made into a crude enzyme

solution to measure hydrolase activity at 24 hours. The last part was stored in a -20 ℃

freezer to characterize microbial community structures via 16S rRNA sequencing.

2.3 Analytical methods

2.3.1 Chemical and enzyme analysis

The following physicochemical properties and enzyme activity were measured

during the aerobic fermentation process using the methods described above (Zhang et

al., 2011; Loh and Ting, 2016): pH, temperature (T), moisture content (MC), total

nitrogen content (TN), organic matter content (OMC), carbon-nitrogen ratio (C/N), and

protease, amylase, lipase, and cellulase activity. All samples were analyzed in triplicate.

Samples were oven-dried at 105 ℃ to determine moisture content (MB90 moisture

meter, America). The oven-dried samples were further heated at 550 ℃ for 5 h for the

determination of OMC. Total organic carbon content (TOC) was calculated by dividing

OMC by 1.8 (Tweib et al., 2014). TN was determined via Kjeldahl nitrogen

determination. C/N was determined as the ratio of TN to TOC. pH was determined

using a pH meter (pH700, Eutech, China). The temperature was measured using a

five-point method, whereby thermometers were inserted around the feedstock and at the

center point. Protease activity was determined using the Folin’s phenol reagent method,

amylase activity was measured using the starch-iodine colorimetric method, lipase

activity was analyzed with the alkali colorimetric method, and cellulase activity was

determined with the 3, 5-dinitrosalicylic acid method.

2.3.2 Microbial community analysis

To determine the effect of different food waste compositions on microbial

community structure during aerobic fermentation, five samples were periodically

collected from each reactor, sample A on day 8, sample B on day 13, sample C on day

18, sample D on day 23, sample E on day 28. A1-E1 were designated as R1 (formulated

diet), A2-E2 as R2 (high oil/fat and salt diet), A3-E3 as R3 (high oil/fat and sugar diet),

and A4-E4 as R4 (vegetarian diet). After DNA extraction, the V4 region of the bacterial

16S rRNA genes was amplified using primers 515F (5’- GTG CCA GCM GCC GCG

GTA A -3’) and 806R (5’- GGA CTA CHV GGG TWT CTA AT -3’). Afterward, the

amplification products were purified and submitted for sequencing (Ion S5TMXL

platform; Novogene, Beijing, China). The sequences were clustered into operational

taxonomic units (OTUs) at a 97% similarity threshold using the UPARSE pipeline

(version 7.0.1001). OTUs were classified taxonomically using the SILVA database

(version 132). The tag sequence with the highest abundance was selected as a

representative sequence within each cluster. Between-groups Venn analysis was

performed with the R programming language (version 3.4.1) to identify unique and

common OTUs (Edgar et al., 2013). The alpha diversity of each sample was determined

as with the “ACE,” “Chao1,” “Shannon,” and “Simpson” diversity indexes. Rarefaction

curves were calculated and plotted with the R programming language to compare the

bacterial OTU diversity between different samples (Zhang et al., 2019). The relationship

between flora and environmental factors were analyzed with distance-based redundancy

analysis (dbRDA) and Spearman correlation analysis. Taxonomic composition analysis

was used to characterize microbial community structures.

All sequences derived from this study have been deposited in the National Center

for Biotechnology Information Sequence Read Archive under the accession number

SRP250398.

2.4 Statistical analysis

Analysis of Variance (ANOVA) was conducted using the SPSS 22.0 statistical

software package. Repeated measures ANOVA was used to identify significant

differences between the replicate trial results. All results were deemed statistically

significant when P-values were < 0.05.

3. Results and discussion

3.1 Physicochemical properties

Physicochemical properties are important factors affecting the aerobic fermentation

process of food waste. Therefore, in this study, the time-dependent changes in

temperature, moisture content, pH, total nitrogen content, and organic matter content of

different food waste were measured.

Temperature (T) is an important factor that affects the aerobic fermentation rate

and microbial physiological activity of food waste. Excessively high temperatures can

lead to widespread microbial death and serious effects on the hydrolase system.

Conversely, if the temperature is too low, the growth and metabolic capacity of

microorganisms will be inhibited, thus affecting the food waste fermentation process

(Liu et al., 2017). Fig. 2a illustrates the time-dependent temperature changes during the

aerobic fermentation process of the four food waste groups. From day 0 to day 8, all

groups exhibited the same increasing temperature trend, whereby the temperature

increased from 29.10 ℃ to the maximum value of 44.77 ℃. In the early stage of the

reaction, the growth and metabolism of microorganisms were vigorous and therefore the

organic matter was consumed and utilized by microorganisms and a large amount of

heat was released (Xi et al., 2015). After the temperature reached 44.5 ℃, a stable state

was maintained. At this time, the temperature of each reactor had no significant change,

and the floating range was between 42.5 ℃ and 44.5 ℃. No significant differences

were observed between R2 and R3 on days 23 and 28 (P = 0.608; P = 0.256), which

may indicate that the composition of R2 and R3 had no significant effect on temperature

during the late fermentation period. Between days 18 and 28, the temperature of R1 was

always the highest. This may be because compared with R2, R3, and R4, R1 contained

more proteinaceous organic matter, which was more readily degraded by

microorganisms, thus producing more heat to maintain the temperature of the system.

During the composting process, the moisture content (MC) not only affects the

metabolism of microorganisms but also determines the food waste aerobic fermentation

efficiency. When the moisture content is too high (i.e., > 70%), it is difficult to maintain

the air circulation inside the reactor, and food waste will also tend to produce acid decay

gas due to anaerobic fermentation. In contrast, excessively low moisture contents inhibit

microbial activity. Fig. 2b illustrates the time-dependent moisture content changes

during the aerobic fermentation process of the four food waste groups. Between days 0

and 13, the moisture content of all four experimental groups exhibited a rapidly

decreasing trend from 57.40%, 44.27%, 53.69%, and 59.27% to the lowest values of

24.43%, 20.33%, 20.87% and 26.73%, respectively. This was likely because the internal

material temperature continued to rise, as well as ventilation and artificial agitation, and

therefore the water evaporation rate was greater than the amount of water that was

generated by the microbial decomposition of organic matter. On day 13, the moisture

content of R2 and R3 was 20.33% and 20.87%, respectively, which was significantly

lower than that of the other two experimental groups (P < 0.05) given that their

moisture content was affected by the fat and oil content in the system. Relevant studies

have shown that fat and oil can inhibit the composting process (Chang and Hsu, 2008;

Maliki and Lai., 2011). This could be because the metabolic enzyme activity of

microorganisms is affected by high fat/oil content and therefore relatively less water is

produced. On day 13 of fermentation, R2 and R3 exhibited visible lumps, and the

moisture content of R1 and R4 also decreased to approximately 25%. When the water

content is too low (i.e., < 15%), microbial activity comes to a halt (Ahn et al., 2008).

Therefore, after day 13, small amounts of distilled water (75 mL) were repeatedly added

to each of the four reaction chambers to improve the moisture content and ensure a

smooth food waste aerobic fermentation process. At the same time, bacterial

metabolism became active, and microbial oxidative decomposition of organic matter

also produced substantial amounts of water. During the fermentation process, the

moisture content of R4 was significantly higher than that of the other groups (P < 0.05),

this was attributed to the higher water content of vegetables, which comprised a large

proportion of R4.

pH is another key factor that affects the aerobic fermentation efficiency of food

waste and is therefore a significant indicator to monitor during food waste fermentation.

The pH variations of the food waste with different properties were relatively consistent

within 28 days. It can be observed from the figure that the pH gradually increased, and

the initial pH was generally between 3.90 and 3.96 (Fig. 2c). When pH < 6.0,

acidophilic microorganism activity was strong, which could easily degrade readily

fermentable organic materials in the system into small organic matter molecules. This

generated organic acids and released substantial amounts of heat, which made it

difficult for the pH of the food waste to rise. With the rise in temperature and the

continuation of the aerobic fermentation process, organic acid gradually volatilized and

organic nitrogen in the materials gradually decomposed under the action of

microorganisms, which entered into the system in the form of ammonia nitrogen.

Consequently, pH values gradually increased (Zhang et al., 2017). The pH of R4 was

the highest during the composting process, whereas the pH of R2 and R3 rose relatively

slowly and exhibited no significant differences (P = 1.000). This was because R4 was

mainly composed of vegetables, which was consistent with a relevant study that found

composting vegetable waste maintained high pH values (Liu et al., 2017; Ghinea and

Leahu., 2020). Moreover, R2 and R3 contained higher fat and oil, which led to the

production of fatty acids in the system. Chang and Hsu (2008) concluded that the

formation of fatty acids after the hydrolysis of fat was the key mechanism of system

acidification. Therefore, in the process of aerobic fermentation, changes in pH were the

result of mutual transformation between different substances.

Nitrogen is an important nutrient element in the aerobic fermentation process of

food waste. Fig. 2d shows the time-dependent changes in the total nitrogen content of

food waste. Between days 0 and 13, the TN contents of R1, R2, and R3 were higher

than that of R4, this was because R4 contained a minimal protein content. There were

no significant differences in the TN changes between R2 and R3 during the early and

late stages of fermentation (P = 0.771; P = 0.158), which may be caused by the similar

composition of R2 and R3. In the process of fermentation, nitrogen-containing organic

matter such as protein was decomposed via microbial activity, and a large amount of

ammonium nitrogen was produced, which made the total nitrogen content begin to rise

in the initial stage of the reaction (Wang et al., 2016). With the gradual volatilization of

ammonium nitrogen from the material, the pH began to rise, and protease activity began

to increase, while the total nitrogen content in the material decreased accordingly

(Zhang et al., 2017). On day 18, the TN of the four groups (R1- R4) was 0.98%, 0.62%,

0.81%, and 0.77%, respectively (Fig. 2d). Among them, the total nitrogen of R1, R2,

and R4 showed an increasing trend. This was because during the later stages of aerobic

fermentation, the mineralization and decomposition of a large number of organic

substances led a reduction in dry matter, and a decrease in nitrogen loss results in

relatively high TN concentrations, thus leading to a gradual increase in TN content

(Bernal et al., 2009). These results were consistent with the observations of previous

studies (Tweib et al., 2014; Ghinea and Leahu, 2020). In contrast, the TN content in R3

exhibited a decreasing trend. This TN decrease in R3 may be related to the changes in

the microbial community structure. Relevant studies identified Paenibacillus as a

nitrogen-fixing bacterium during the composting process, and decreases in their

abundance may result in TN decreases (Attar, 2018; Zhang et al., 2019).

Paenibacillus abundance in R3 exhibited a decreasing trend in the late stage of

fermentation, and was lower than that of the other three groups (Fig. 5b). Therefore, the

TN content of R3 exhibited a decreasing trend on day 18. Except for the 13th day, the

TN content of R1 was the highest. Notably, the protein content of R1 was higher than

that of the other groups, and its organic matter degradation rate was also the fastest,

which resulted in an overall decrease in carbon to nitrogen ratio coupled with an

increased total nitrogen content (Bernal et al., 2009). It can be said that the change in

TN resulted from the concentration effect caused by ammonia volatilization and organic

matter mineralization. Based on the total nitrogen results, the formulated diet (R1) food

waste nitrogen content was the largest, which indirectly indicated that R1 waste

contained more nutrients to sustain microorganism growth and reproduction.

Organic matter content (OMC) is an important indicator of compost maturity, and

its dynamic changes can directly reflect the aerobic fermentation process (Wang et al.,

2019). During the entire fermentation process, the organic matter content of the four

experimental groups declined slowly, from 80.40%, 79.18%, 79.20%, and 77.71% to

71.20%, 70.60%, 71.18%, and 72.66%, respectively (Fig. 2e). This may be due to the

carrier of the compound microbial agent that was used (i.e., mainly wheat straw), which

had a complex structure and more insoluble cellulose, which is difficult to be degraded

by microorganisms and therefore the organic matter degradation of food waste was not

enough to cause the organic matter in the waste heap to change (Wang et al., 2016). A

microbial agent containing rice bran was added to citrus peel compost, after which the

organic matter decline rate was only 10% after 35 days of municipal solid waste

composting (Wang et al., 2019). Another study used cornstalk as a bulking agent in

co-composting sewage sludge and municipal solid waste, which resulted in

approximately 8.5% decreases in organic matter after composting (Zhang et al., 2017).

During early fermentation stages, the water and nutrient content of the reactor remained

sufficient as the system temperature increased, and the growth and metabolism of

thermophilic microorganisms were vigorous. Easily degradable organic matter was

consumed and decomposed first and thus the organic matter content began to decline

and hydrolase activity began to increase. The organic matter content of R1 decreased

the fastest (i.e., by 9.20%), followed by that of R2 and R3, which decreased by 8.57%

and 8.01%. This was likely because the fat content of these materials was more difficult

to degrade than protein and sugar. Compared with R1, R2, and R3, there were

statistically significant differences in the OMC changes of R4 in the early and late

fermentation periods (P < 0.05). The organic matter content in vegetarian waste

declined the least. This was because these waste contained more cellulose, which is one

of the most difficult food waste compounds to degrade (Zhang et al., 2014). From an

organic matter degradation standpoint, the formulated diet (R1) food waste was the

most favorable. Therefore, it is reasonable to speculate that the composting process of

the formulated diet food waste was relatively faster than that of the other dietary

regimes.

The C/N ratio is an important indicator of food waste compost maturity. Fig. 2f

shows the time-dependent changes in the C/N ratio of food waste. From days 0 to 28,

the C/N ratio of the four food waste groups changed from 35.78, 43.19, 43.60, and

98.39 to 31.74, 53.25, 56.80, and 44.85, respectively. Related studies have found that

the C/N ratio in the composting process of food waste generally decreases from an

initial value of approximately 35 to below 25 (Wang et al., 2016; Wang et al., 2017). In

contrast, the C/N ratio in this study was generally high, which may be caused by

differences in the detection methods. The aforementioned studies used a CNS elemental

analyzer to determine C/N ratios, whereas the present study determined C/N ratios by

calculating the ratio of TN to TOC. TN content values in the present study were low

because the total nitrogen determined by the Kjeldahl method does not account for the

nitrate value, and therefore the C/N ratio of the present study was higher (Zhang et al.,

2011). The C/N ratio of R1 decreased from the beginning to the end of the experiment

during the aerobic fermentation process, whereas the C/N ratio of R2 and R3 increased

significantly. This indicated that the aerobic fermentation process of R2 and R3 was

affected by its components, which led to pronounced C/N ratio changes. This may have

been because the microbial community structure of high oil/fat, high sugar, and high

salt food waste was different from that of the formulated diet food waste. High C/N

ratios are closely related to changes in microbial community structure, which was

supported by the microbial community composition data determined herein (Fig. 5a).

The Thermoactinomyces and Bacillus abundance of R2 and R3 was higher than that of

R1. Zhang et al. (2020) found that Thermoactinomyces, Bacillus, and Pseudomonas

maintained high bacterial abundance in compost with high C/N ratios. The C/N ratio of

R4 showed a significant decrease compared to the initial stage of aerobic fermentation,

which may be due to the relatively low total nitrogen content and higher organic matter

content of R4 during aerobic fermentation.

3.2 Hydrolase activity

Hydrolase activity is an important indicator for food waste aerobic fermentation

monitoring. Changes in substrate hydrolase activity during different fermentation stages

can reveal the process of mineralization and decomposition of organic matter in food

waste from a microbial structure standpoint (Awasthi et al., 2017). During the early

stage of fermentation, the abundance of fermentable materials (i.e., substrates) and the

large number of reproducing microorganisms translates to a large concentration of

hydrolytic enzymes that could degrade organic substances, which could effectively

degrade complex macromolecular organic matters in food waste. During the

post-fermentation period, the microorganisms were restricted by the internal

environmental factors of the system, their metabolic activities began to decelerate, and

the activity of substrate degrading enzymes began to decline, which indicated that the

whole aerobic fermentation process was coming to an end.

Variations in hydrolase (protease, amylase, lipase, cellulase) activity were

examined to identify microbial dynamics for various organic waste degradation

processes. In this study, the activity of protease, amylase, and lipase increased first and

then decreased, whereas the activity of cellulase remained at a very low level. Moreover,

significant differences were observed in the variation of the activity of the four

hydrolases examined herein between different food waste compositions (P < 0.05),

which indicated that the enzyme activity characteristics of each food composition were

significantly different, which was closely linked to the different microbial flora

characteristics associated with food compositions.

Fig. 3a illustrates the time-dependent protease activity changes during the aerobic

fermentation of food waste. The protease activity of the four food waste groups

increased at first and then decreased. The protease activity of the R4 reached a

maximum value of 70.05 U on day 13. After 18 days, the enzyme activity of R1, R2,

and R3 reached their maximum values of 116.70, 100.06, and 104.67 U, respectively.

However, during this period, there were no significant differences in the protease

activity of R2 and R3 (P = 0.459; P = 0.053), which may be because the salt and sugar

contents in the food waste examined herein did not significantly affect microbial

protease secretion. A high protease activity indicates that the microorganisms consume

and degrade the protein organics quickly. During the early stage of fermentation, with

the increase of material temperature, the content of protein organic matter in the reactor

was sufficient, and the microbial metabolism and growth capacity was strong, thus

secreting a large number of protease substances. Due to the high protein content in R1,

R2, and R3, the protease activity in the system was higher than the protease activity in

R4. In the middle and late stages of fermentation, the metabolic activity of

microorganisms gradually decreased, and the protease activity of the four experimental

groups also decreased after reaching a peak.

Fig. 3b illustrates the time-dependent amylase activity changes during the aerobic

fermentation of food waste. The amylase activity of the four groups increased at first

and then decreased. The amylase activity of R1 and R3 reached its maximum value on

the 8th day at 733.30 and 666.67 U, respectively. On day 13, the amylase activity of R2

and R4 reached its maximum value at 400.00 and 533.33 U, respectively. During

aerobic fermentation, starch and saccharides were first consumed and decomposed by

microorganisms when the materials contained a substantial proportion of rice, which

could provide sufficient energy for amylase-secreting microorganisms. Therefore,

amylase activity in the four food waste groups remained at a high level. Moreover, oil

and salt are known to inhibit the fermentation process, which in turn can affect the

hydrolase activity (Chang and Hsu, 2008; Chen et al., 2008; Maliki and Lai., 2011).

Therefore, the amylase activity of R1, R3, and R4 food waste was higher than R2.

Additionally, there were significant differences in the amylase activities of R2 and R3

during the fermentation period, which may be due to the high salt content in R2

affecting the amylase activities in the system. These results indicated that the effect of

sugar on amylase activity is less than that of salt.

Fig. 3c illustrates the time-dependent changes in food waste lipase activity. During

aerobic fermentation, there was no significant difference in lipase activity between R2

and R3 (P = 0.730). This could be due to the similar oil composition in R2 and R3.

During the fermentation process, the lipase activity of the four food waste groups

increased at first and then decreased. Due to the low fatty acid content in the materials,

fewer nutrients were required to induce lipase secretion from microorganisms.

Therefore, the lipase activity in the four experimental groups remained low. The lipase

activity of R2, R3, and R4 reached the maximum values of 4.13, 6.03, and 1.55 U on

day 13, respectively, whereas the lipase activity of R1 reached its maximum value of

4.10 U on day 18. Oil and fat are known to be more difficult to decompose than

carbohydrates and proteins (Chang and Hsu, 2008). The starch and protein organic

contents of R1 were more abundant than in the other treatments. Such organics were

used first in the process of organic degradation, whereas small amounts of fatty

substances that were initially difficult to degrade were consumed and decomposed later,

and therefore the maximum lipase activity of R1 manifested later than in the other three

experimental groups.

Fig. 3d illustrates the time-dependent changes in food waste cellulase activity.

Notably, the changes in cellulase activity in the four experimental groups remained

largely constant throughout the entire process, and the enzyme activity remained low.

The maximum cellulase activity of the R1, R2, R3, and R4 groups was 0.01, 0.01,

0.0089, and 0.130 U, respectively. This was because the cellulase was prone to denature

and inactivate under strongly acidic conditions, and the optimal pH for cellulase

produced by common fungi ranges from neutral to slightly alkaline. Moreover, the

cellulose degradation process is much more complex than that of other organic materials

such as sugars and proteins (Awasthi et al., 2017). The changes in cellulase activity

were consistent with the changes in pH value of food waste in the four groups (Fig. 2c).

On day 23, the pH value of R1 and R4 increased slightly and, at the same time, cellulase

activity also showed a slowly increasing trend (Fig. 3d). On the other hand, the

composite microbial agent carrier used in the experiment was mainly wheat straw,

which contains substantial amounts of refractory cellulose, and therefore the cellulase

activity of the four food waste groups remained low.

Overall, the amylase and protease activity of R1 was higher than that of other

treatments due to its primary protein and starch composition. From R2 and R3

hydrolase activity results, the influence of salt and sugar on amylase activity was

significant, but exerted no significant effects on protease activity. Moreover, the

cellulase activity of R4 was higher than that of the other three treatments.

3.3 Microbial community structure

3.3.1 OTUs clustering analysis

The total number of OTUs and the number of shared OTUs among the four food

waste groups (R1, R2, R3, and R4) and OTU changes within each group can be easily

observed from the Venn diagram (Fig. 4). Notably, it can be observed that 6,446 OTUs

were commonly shared between R1, R2, R3, and R4, which accounted for a significant

proportion of the total OTUs. Given that the system was small and the temperature and

pH changes did not vary considerably throughout the entire reaction, the overall

community changes remained relatively consistent. These results indicated that the

number of OTUs in the aerobic fermentation process of the four food waste groups was

similar. However, the number of unique OTUs in the R3 group was the largest (i.e., 21).

The independent OTUs of R3 decreased sharply compared with the other three groups,

indicating that the material properties in R3 had a great influence on community

composition (Fig. 4c). All of these observations suggest that the microbial diversity in

the R3 group was relatively higher than that in the R1, R2, and R4 groups. According to

the lipase activity results (Section 3.2), the effect of salt on lipase activity was greater

than that of sugar, whereas the OTU cluster analysis results demonstrated that the effect

of sugar on community composition was greater than that of salt. Some studies have

reported that both sugar and salt can affect microbial activity. For instance, Li et al.

(2019) investigated the effect of salt content on the anaerobic fermentation of food

waste and reported that high salt content could inhibit the metabolic activity of

microorganisms. Moreover, Medeiros et al. (2006) found that fungi abundance would

increase in response to high sugar content. Therefore, both salt and sugar contents have

different effects on the physiological and biochemical indexes of microorganisms. The

adverse effect of salt on physicochemical properties such as water content and pH is

more significant than that of sugar (Li et al., 2019). Therefore, different food waste

components in the waste disposal process are worth further discussion. R1 exhibited the

highest number of shared OTUs, showing that this diet had the least effects on food

waste-associated community changes.

3.3.2 Microbial community diversity

To investigate the diversity of microbial communities during the aerobic

fermentation of different food waste compositions, the alpha diversity analysis index

(Shannon, Simpson, Chao1, ACE) of different samples at a 97% consistency threshold

was statistically analyzed. The Chao1 and ACE indices can be used to calculate

community richness while the Shannon and Simpson indices are usually used to reflect

the diversity of microbial communities. The Chao1 of R2 and R3 were 258.73 and

272.21, respectively, and the ACE indices of R2 and R3 were 269.69 and 281.42,

respectively (Table 2), which were all higher than those of the R1 and R4 groups. In

contrast, the Shannon and Simpson indices of the R1 and R4 groups were higher than

those of R2 and R3. All of these data indicated that the species richness of R2 and R3

was higher than that of R1 and R4, whereas the community diversity of R1 and R4 was

higher than that of R2 and R3. Generally, more astringent environmental conditions

lead to a simpler community structure (Vasconcelos et al., 2016). After the elimination

of various adverse environmental factors, only microorganisms with a strong tolerance

can survive. In these cases, the growth of a single species is typically more vigorous,

thus resulting in a simpler whole-community diversity. Kitamura et al. (2015) found

that a food waste community contained a large number of Firmicutes, which inhibited

the growth of other bacteria, thereby reducing community diversity. Despite oxygen

limitations, Partanen et al. (2010) reported high temperatures during the

high-temperature composting stage, which was coupled with a low species richness,

where Actinomycetes and Thermoactinomyces were predominant.

3.3.3 Taxonomic composition analysis

Fig. 5 illustrates the genus-level distribution of the top 30 microbial communities

in the four food waste groups. The dominant bacteria in the food waste aerobic

fermentation process included Bacillus, Thermoactinomyces, Paenibacillus, Cohnella,

and others. All of these are known to secrete hydrolase during the fermentation process,

thereby effectively promoting the decomposition and transformation of food waste

organic matter (Ke et al., 2010; Rai et al., 2010; Wushke et al., 2013; Msarah et al.,

2020). The abundance of Bacillus was the highest in all the dietary groups examined

(i.e., 37.60, 48.24, 40.55, and 40.08%, respectively) (Fig. 5a). Upon further analysis of

the Bacillus classification level, it was found that the main Bacillus species of the four

food waste groups was Bacillus coagulans, with an abundance of more than 20%.

Related studies have shown that Bacillus coagulans is highly acid-resistant and can

decompose sugars to produce L-lactic acid at 45~50 ℃. Moreover, it can also secrete

large amounts of protease and amylase during the fermentation process to promote the

decomposition and transformation of macromolecular organic matter (Msarah et al.,

2020). The relative abundance of Thermoactinomyces in the four food waste groups was

18.69, 34.45, 27.56, and 23.83%, respectively. These organisms exhibit heat resistance

and are known to survive in harsh environments, metabolizing at 45 ℃ and producing a

variety of heat-resistant enzymes, such as amylase, serine, protease, and lipase (Ke et al.,

2010). The presence of these bacteria can effectively promote the aerobic fermentation

process of food waste, which was consistent with the aforementioned results that

determined that the enzyme activity remained high during the aerobic fermentation of

food waste. Additionally, the relative abundance of Paenibacillus in the R1 group was

11.73%, which was higher than that of the R2 and R3 groups. Paenibacillus is able to

secrete more proteases (Rai et al., 2010), which is consistent with the previously

reported high protease activity in food waste from formulated diets. At the genus level,

there were a large number of Bacillus and Thermoactinomyces in R2 and R3, which

accounted for more than 70% of the total bacterial flora, this was more than in R1 and

R4, indicating that these bacteria may have a certain preference for oil/fat, salt, sugar,

and other food waste. Among these classifications, Thermoactinomyces are known to be

heat-resistant and quite suitable for survival in adverse environmental conditions, which

might have inhibited the growth of other bacteria, thereby decreasing community

diversity (Ke et al., 2010; Kitamura et al., 2015). Specifically, the relative abundance of

Cohnella in R4 was the highest, at 4.91%. This bacterium played an important role in

degrading cellulosic substances, which increased cellulase activity during the later stage

of fermentation (Wushke et al., 2013).

Fig. 5b illustrates the intra-group abundance of species at the genus level during

the aerobic fermentation of different food waste. Although the food waste components

were different, four types of bacteria (Bacillus, Thermoactinomyces, Paenibacillus, and

Cohnella) occupied a large proportion in almost every fermentation period. These

bacteria also showed a tendency to increase first and then decrease during the

fermentation process. Such results indirectly revealed the dynamic biochemical

degradation process of food waste, which was consistent with the above-mentioned

changes in hydrolase activity.

3.3.4 Environmental factors correlation analysis

Distance-based redundancy analysis (dbRDA) is typically used to reflect the

relationship between the quantitative characteristics and environmental factors of

different sample species compositions (Warton et al., 2012). In this study, dbRDA was

used to reveal the relationship between microbial community structure composition and

environmental factors during food waste aerobic fermentation. dbRDA1 and dbRDA2

explained 48.90% and 29.72% of the total variation, respectively (Fig. 6a). The length

of the arrow line represents the correlation between pH, T, M (moisture content), and

community distribution, and the angle between the arrow line and the sorting axis

reflects the correlation between environmental factors and the sequencing axis. Longer

lengths indicate a stronger correlation. When the angle between environmental factors is

smaller, this demonstrates a positive correlation between the two. It can be observed

that pH was negatively correlated with T, M (P < 0.05), and there is a significant

correlation between T, pH, M and community distribution (Fig. 6a). Such results were

consistent with a previous study (Zhang et al., 2011; Awasthi et al., 2018; Sun et al.,

2020). Moreover, there was a significant positive correlation between pH and

community distribution in R1 and R4, indicating that R1 and R4 microbial communities

were more sensitive to pH variation than R2 and R3. Moreover, R2 and R3 were also

insensitive to T and M due to their characteristically high oil/fat and salt/sugar content.

Unlike pH, the influence of T on different food waste community changes was not

significant. Given that the experimental temperature was largely maintained at

approximately 45 ℃, the dominant bacteria of different food waste did not change

significantly.

Spearman rank correlation is mainly used to demonstrate the relationship between

environmental factors and the abundance of microbial species during aerobic

fermentation. Correlation and significance P values between pairs were obtained to

explore the relationship between environmental factors and species. There was a certain

correlation between three environmental factors (T, pH, and M) at the genus level for a

total of 35 bacterial species (Fig. 6b). Moreover, temperature had a positive correlation

with a total of 28 bacterial species, whereas pH and moisture content had a negative

correlation with most bacteria. Brevibacillus had a significant positive correlation with

temperature, which was consistent with literature that reported that Brevibacillus has the

ability to degrade cellulose at high temperatures (Maeda et al., 2011). The four bacteria

(Treponema, Truepera, Bellilinea, and Azotobacter) exhibited a significant inverse

relationship with pH. Among them, Bellilinea is known to be able to use a variety of

carbohydrates as a carbon source in acidic and anaerobic environments (Yamada et al.,

2007). Alcaligenes had a significant positive correlation with pH and moisture content

(P < 0.05), which was consistent with the fact that Alcaligenes cannot easily survive in

low pH and high salinity environments (Kim et al., 2019)

3.4 Evaluation and suggestions

This study characterized the biochemical degradation of food waste under four

different dietary regimes. The formulated diet food waste exhibited a rapid decline in

organic matter and a high hydrolytic enzyme activity, whereas the high oil/fat and salt

food waste exhibited low pH, and its high salt content affected hydrolase activity during

aerobic fermentation. Compared with salt, the negative impact of sugar on

physicochemical properties was much smaller. Therefore, future studies should focus on

the removal of oil/fat and salt from food waste during the composting process.

Vegetarian waste had relatively stable physicochemical properties and enzyme activity

during aerobic fermentation. However, the negative effects of high water content and

the slow degradation of organic matter should be considered in vegetarian waste

composting processes.

Current laboratory-scale study results suggest that improving diet quality was an

important factor that affected food waste aerobic composting. Therefore, diet quality

should be improved by avoiding high oil/fat, high sugar, and high salt diets. At the same

time, the above-described information can help individuals and organizations to

implement in-situ food waste treatment as much as possible. For instance, oil and salt in

waste should be pre-filtered or removed during the aerobic fermentation process of

household food waste to facilitate food waste transportation and centralized treatment.

4. Conclusions

In the process of aerobic fermentation, the physicochemical properties and

hydrolase activity of different food waste were found to be significantly different. The

dominant bacteria of the four food waste were Bacillus, Thermoactinomyces,

Paenibacillus, and Cohnella. The species richness of high oil/fat and salt and high

oil/fat and sugar food waste was higher than formulated and vegetarian diet waste,

whereas the community diversity of formulated and vegetarian diet food waste were

higher than that of high oil/fat and salt and high oil/fat and sugar waste. Taken together,

these results are conducive to improving the fermentation efficiency of food waste

compost.

Acknowledgments

This research was supported by the National Key R&D Program of China (Grant

No. 2018YFC1901000), the Strategic Priority Research Program of the Chinese

Academy of Sciences (Grant No. XDA23040302), the Natural Science Foundation of

China (Grant No. 51678552) and Xiamen WES&WQGE (201902). The authors also

thank mjeditor (www.mjeditor.com) for their assistance in English language editing.

References

1. Ahn, H.K., Richard, T.L., Glanville, T.D., 2008. Optimum moisture levels for biodegradation of

mortality composting envelope materials. Waste Manag. 28 (8), 1411-1416.

2. Attar, A., 2018. Effects of Azorhizophilus paspali and Paenibacillus mucilaginosus as

biofertilizer and determination of nutritional efficiency by sensors. Arab J Sci Eng. 43 (7),

3477-3484.

3. Awasthi, M.K., Wong, J.W.C., Kumar, S., Awasthi, S.K., Wang, Q., Wang, M., Ren, X., Zhao, J.,

Chen, H., Zhang, Z., 2017. Biodegradation of food waste using microbial cultures producing

thermostable α-amylase and cellulase under different pH and temperature. Bioresour.

Technol. 248, 160-170.

4. Awasthi, S.K., Wong, J.W.C., Li, J., Wang, Q., Zhang, Z., Kumar, S., Awasthi, M.K., 2018.

Evaluation of microbial dynamics during post-consumption food waste composting.

Bioresour. Technol. 251, 181-188.

5. Bernal, M.P., Alburquerque, J.A., Moral, R., 2009. Composting of animal manures and chemical

criteria for compost maturity assessment. A review. Bioresour. Technol. 100 (22),

5444-5453.

6. Brigham, E.P., Kolahdooz, F., Hansel, N., Breysse, P.N., Davis, M., Sharma, S., Matsui, E.C.,

Diette, G., McCormack, M.C., 2015. Association between Western diet pattern and adult

asthma: a focused review. Ann Allergy Asthma Immunol. 114 (4), 273-280.

7. Chang, J.I., Hsu, T.E., 2008. Effects of compositions on food waste composting. Bioresour.

Technol. 99 (17), 8068-8074.

8. Chen, H., Jiang, W., Yang, Y., Yang, Y., Man, X., 2017. State of the art on food waste research:

a bibliometrics study from 1997 to 2014. J. Clean. Prod. 140, 840-846.

9. Chen, Y., Cheng, J.J., Creamer, K.S., 2008. Inhibition of anaerobic digestion process: a review.

Bioresour. Technol. 99 (10), 4044-4064.

10. Conrad, Z., Niles, M.T., Neher, D.A., Roy, E.D., Tichenor, N.E., Jahns, L., 2018. Relationship

between food waste, diet quality, and environmental sustainability. PLoS One. 13 (4),

e0195405.

11. Edgar, R.C., 2013. UPARSE: highly accurate OTU sequences from microbial amplicon reads.

Nat. Methods. 10 (10), 996-998.

12. Food and Agriculture Organization of the United Nations. 2016. Plates, pyramids, planet.

Developments in national and sustainable dietary guidelines: a state of play assessment.

http://www.fao.org/sustainable-food-value-chains/library/details/en/c/415611/.

13. Ghinea, C., Leahu, A., 2020. Monitoring of fruit and vegetable waste composting process:

relationship between microorganisms and physico-chemical parameters. Processes. 8 (3),

302.

14. Kastorini, C.-M., Milionis, H.J., Ioannidi, A., Kalantzi, K., Nikolaou, V., Vemmos, K.N.,

Goudevenos, J.A., Panagiotakos, D.B., 2011. Adherence to the Mediterranean diet in relation

to acute coronary syndrome or stroke nonfatal events: A comparative analysis of a

case/case-control study. Am. Heart J. 162 (4), 717-724.

15. Ke, G.R., Lai, C.M., Liu, Y.Y., Yang, S.S., 2010. Inoculation of food waste with the

thermo-tolerant lipolytic actinomycete Thermoactinomyces vulgaris A31 and maturity

evaluation of the compost. Bioresour. Technol. 101 (19), 7424-7431.

16. Key, T.J., Appleby, P.N., Rosell, M.S., 2006. Health effects of vegetarian and vegan diets. Proc

Nutr Soc. 65 (1), 35-41.

17. Kim, M., Abdulazeez, M., Haroun, B.M., Nakhla, G., Keleman, M., 2019. Microbial

communities in co-digestion of food wastes and wastewater biosolids. Bioresour. Technol.

289, 121580.

18. Kitamura, R., Ishii, K., Maeda, I., Kozaki, T., Iwabuchi, K., Saito, T., 2015. Evaluation of

bacterial communities by bacteriome analysis targeting 16S rRNA genes and quantitative

analysis of ammonia monooxygenase gene in different types of compost. J. Biosci. Bioeng.

121 (1), 57-65.

19. Li, J., Wang, X., Cong, C., Wan, L., Xu, Y., Li, X., Hou, F., Wu, Y., Wang, L.,

2020. Inoculation of cattle manure with microbial agents increases efficiency and promotes

maturity in composting. 3 Biotech. 10 (3).

20. Li, X., Huang, J., Liu, Y., Huang, T., Maurer, C., Kranert, M., 2019. Effects of Salt on Anaerobic

Digestion of Food Waste with Different Component Characteristics and Fermentation

Concentrations. Energies. 12 (18), 3571.

21. Liu, L., Wang, S., Guo, X., Zhao, T., Zhang, B., 2017. Succession and diversity of

microorganisms and their association with physicochemical properties during green waste

thermophilic composting. Waste Manag. 73, 101-112.

22. Loh, J.Y., Ting, A.S.Y., 2016. Effects of potential probiotic Lactococcus lactis subsp. lactis on

digestive enzymatic activities of live feed Artemia franciscana. Aquacult Int. 24 (5),

1341-1351.

23. Maeda, T., Yoshimura, T., Garcia-Contreras, R., Ogawa, H.I., 2011. Purification and

characterization of a serine protease secreted by Brevibacillus sp. KH3 for reducing waste

activated sludge and biofilm formation. Bioresour. Technol. 102 (22), 10650-10656.

24. Maliki, A.D., Lai, K.M., 2011. Design and application of a pre-composting test step to determine

the effect of high fat food wastes on an industrial scale in-vessel composting system. Int.

Biodeter. Biodegr. 65 (6), 906-911.

25. Medeiros, P.M., Fernandes, M.F., Dick, R.P., Simoneit, B.R., 2006. Seasonal variations in sugar

contents and microbial community in a ryegrass soil. Chemosphere. 65 (5), 832-839.

26. Msarah, M.J., Ibrahim, I., Hamid, A.A., Aqma, W.S., 2020. Optimisation and production of

alpha amylase from thermophilic Bacillus spp. and its application in food waste

biodegradation. Heliyon. 6 (6), e04183.

27. Nakasaki, K., Hirai, H., Mimoto, H., Quyen, T.N.M., Koyama, M., Takeda, K., 2019. Succession

of microbial community during vigorous organic matter degradation in the primary

fermentation stage of food waste composting. Sci. Total Environ. 671, 1237-1244.

28. Onwosi, C.O., Igbokwe, V.C., Odimba, J.N., Eke, I.E., Nwankwoala, M.O., Iroh, I.N., Ezeogu,

L.I., 2017. Composting technology in waste stabilization: on the methods, challenges and

future prospects. J. Environ. Manage. 190, 140-157.

29. Partanen, P., Hultman, J., Paulin, L., Auvinen, P., Romantschuk, M., 2010. Bacterial diversity at

different stages of the composting process. BMC Microbiol. 10 (1), 94.

30. Petracchini, F., Liotta, F., Paolini, V., Perilli, M., Cerioni, D., Gallucci, F., Carnevale, M.,

Bencini, A., 2017. A novel pilot scale multistage semidry anaerobic digestion reactor to treat

food waste and cow manure. Int J Environ Sci Technol (Tehran). 15 (9), 1999-2008.

31. Qureshi, A.S., Khushk, I., Naqvi, S.R., Simiar, A.A., Ali, C.H., Naqvi, M., Danish, M., Ahmed,

A., Majeed, H., Jatt, A.N.M., Rehan, M., Nizami, A.S., 2017. Fruit waste to energy through

open fermentation. Energy Procedia. 142, 904-909.

32. Rai, S.K., Roy, J.K., Mukherjee, A.K., 2010. Characterisation of a detergent-stable alkaline

protease from a novel thermophilic strain Paenibacillus tezpurensis sp nov AS-S24-II. Appl.

Microbiol. Biotechnol. 85 (5), 1437-1450.

33. Sun, L., Han, X., Li, J., Zhao, Z., Liu, Y., Xi, Q., Guo, X., Gun, S., 2020. Microbial community

and its association with physicochemical factors during compost bedding for dairy cows.

Front. Microbiol. 11.

34. Thyberg, K.L., Tonjes, D.J., 2016. Drivers of food waste and their implications for sustainable

policy development. Resour Conserv Recycl. 106, 110-123.

35. Tweib, S.A.K., Rahman, R.A., Khalil, M.S., 2014. Physicochemical changes in co-composting

process of palm oil mill sludge (POMS) and solid waste (Kitchen Waste) using bin

composter. Arab J Sci Eng. 39 (4), 2455-2462.

36. Vasconcelos, E.A.F., Leitão, R.C., Santaella, S.T., 2016. Factors that affect bacterial ecology in

hydrogen-producing anaerobic reactors. Bioenerg. Res. 9 (4), 1260-1271.

37. Wang, J., Liu, Z., Xia, J., Chen, Y., 2019. Effect of microbial inoculation on physicochemical

properties and bacterial community structure of citrus peel composting. Bioresour. Technol.

291, 121843.

38. Wang, X., Pan, S., Zhang, Z., Lin, X., Zhang, Y., Chen, S., 2017. Effects of the feeding ratio of

food waste on fed-batch aerobic composting and its microbial community. Bioresour.

Technol. 224, 397-404.

39. Wang, X., Zhang, W., Gu, J., Gao, H., Qin, Q., 2016. Effects of different bulking agents on the

maturity, enzymatic activity, and microbial community functional diversity of kitchen waste

compost. Environ Technol. 37 (20), 2555-2563.

40. Warton, D.I., Wright, S.T., Wang, Y., 2012. Distance-based multivariate analyses confound

location and dispersion effects. Methods Ecol. Evol. 3 (1), 89-101.

41. Wushke, S., Levin, D.B., Cicek, N., Sparling, R., 2013. Characterization of enriched aerotolerant

cellulose-degrading communities for biofuels production using differing selection pressures

and inoculum sources. Can J Microbiol. 59 (10), 679-683.

42. Xi, B., He, X., Dang, Q., Yang, T., Li, M., Wang, X., Li, D., Tang, J., 2015. Effect of multi-stage

inoculation on the bacterial and fungal community structure during organic municipal solid

wastes composting. Bioresour. Technol. 196, 399-405.

43. Yamada, T., Imachi, H., Ohashi, A., Harada, H., Hanada, S., Kamagata, Y., Sekiguchi, Y., 2007.

Bellilinea caldifistulae gen. nov., sp. nov. and Longilinea arvoryzae gen. nov., sp. nov.,

strictly anaerobic, filamentous bacteria of the phylum Chloroflexi isolated from

methanogenic propionate-degrading consortia. Int J Syst Evol Microbiol. 57 (10), 2299-2306.

44. Zhang, D., Luo, W., Li, Y., Wang, G., Li, G., 2017. Performance of co-composting sewage

sludge and organic fraction of municipal solid waste at different proportions. Bioresour.

Technol. 250, 853-859.

45. Zhang, J., Zeng, G., Chen, Y., Liang, J., Zhang, C., Huang, B., Sun, W., Chen, M., Yu, M.,

Huang, H., Zhu, Y., 2014. Phanerochaete chrysosporium inoculation shapes the indigenous

fungal communities during agricultural waste composting. Biodegradation. 25 (5), 669-680.

46. Zhang, J., Zeng, G., Chen, Y., Yu, M., Yu, Z., Li, H., Yu, Y., Huang, H., 2011. Effects of

physico-chemical parameters on the bacterial and fungal communities during agricultural

waste composting. Bioresour. Technol. 102 (3), 2950-2956.

47. Zhang, L., Loh, K.-C., Lim, J.W., Zhang, J., 2019. Bioinformatics analysis of metagenomics data

of biogas-producing microbial communities in anaerobic digesters: A review. Renew. Sust.

Energ. Rev. 100, 110-126.

48. Zhang, W., Yu, C., Wang, X., Hai, L., 2019. Increased abundance of nitrogen transforming

bacteria by higher C/N ratio reduces the total losses of N and C in chicken manure and corn

stover mix composting. Bioresour. Technol. 297.

Figure captions:

Fig. 1 Food waste aerobic fermentation reactor

Fig.2. Time-dependent changes in physicochemical properties (T, MC, pH, TN, OMC,

C/N) during the aerobic fermentation of the four food waste groups examined herein.

Fig.3. Changes in hydrolytic enzyme (protease, amylase, lipase, and cellulase) activity

over time during the aerobic fermentation of the four food waste groups examined

herein.

Fig.4. (a) OTU changes within the R1 group. (b) OTU changes within the R2 group. (c)

OTU changes within the R3 group. (d) OTU changes within the R4 group. (e) Total

number of OTUs and number of shared OTUs between the four food waste groups.

Fig. 5. Relative inter-group (a) and intra-group (b) species abundance at the genus level

during the aerobic fermentation of different food waste (samples A1, B1, C1, D1, E1

correspond to the R1 group; samples A2, B2, C2, D2, E2 correspond to the R2 group;

samples A3, B3, C3, D3, E3 correspond to the R3 group; samples A4, B4, C4, D4, E4

correspond to the R4 group).

Fig. 6 (a) dbRDA analysis of the relationship between distances based on flora and

environmental factors. (b) Spearman correlation analysis of the relationship between

specific genera and environmental factors.

Table 1 Properties and proportions of the feedstocks

Formulated diet High oil/fat and salt High oil/fat and sugar Vegetarian diet

Rice 363 g 363 g 363 g 363 g

Lean meat 100 g 0 g 0 g 0 g

Fat meat 0 g 100 g 100 g 0 g

Cabbage 250 g 250 g 250 g 250 g

Celery 0 g 0 g 0 g 100 g

Spice powder 0.35 g 0.35 g 0.35 g 0.35 g

Salt 3 g 12 g 3 g 3 g

Sugar 3 g 3 g 12 g 3 g

Water 150 mL 150 mL 150 mL 150 mL

Oil 3 mL 6 mL 6 mL 3 mL

Microbial agent 80 g 80 g 80 g 80 g

Table 2 Alpha diversity indexes of each group

Community diversity Community richnessGroups

Shannon Simpson Chao1 ACE

R1 4.303 0.884 224.619 229.625

R2 3.879 0.859 258.734 269.689

R3 3.819 0.811 272.214 281.424

R4 4.351 0.902 251.246 248.896

Fig. 1 Food waste aerobic fermentation reactor

Fig.2. Time-dependent changes in physicochemical properties (T, MC, pH, TN, OMC,

C/N) during the aerobic fermentation of the four food waste groups examined herein.

Fig.3. Changes in hydrolytic enzyme (protease, amylase, lipase, and cellulase) activity

over time during the aerobic fermentation of the four food waste groups examined

herein.

Fig.4. (a) OTU changes within the R1 group. (b) OTU changes within the R2 group. (c)

OTU changes within the R3 group. (d) OTU changes within the R4 group. (e) Total

number of OTUs and number of shared OTUs between the four food waste groups.

Fig. 5. Relative inter-group (a) and intra-group (b) species abundance at the genus level

during the aerobic fermentation of different food waste (samples A1, B1, C1, D1, E1

correspond to the R1 group; samples A2, B2, C2, D2, E2 correspond to the R2 group;

samples A3, B3, C3, D3, E3 correspond to the R3 group; samples A4, B4, C4, D4, E4

correspond to the R4 group).

Fig. 6 (a) dbRDA analysis of the relationship between distances based on flora and

environmental factors. (b) Spearman correlation analysis of the relationship between

specific genera and environmental factors.

CRediT Authorship Contribution Statement

Yanzeng Li: Conceptualization, Data curation, Formal analysis, Investigation, Methodology,

Writing - original draft, Writing - review & editing.

Zhou Chen: Data curation, Formal analysis, Investigation, Methodology, Writing - review &

editing.

Yanyan Peng: Formal analysis, Investigation, Methodology.

Kaiming Zheng: Data curation, Formal analysis.

Chengsong Ye: Data curation, Methodology.

Kun Wan: Formal analysis, Methodology.

Shenghua Zhang*: Funding acquisition, Writing - review & editing, Supervision, Review.

Highlights：

Biochemical degradation mechanism of different food waste was proposed.

The physicochemical properties of different food waste were significantly different.

The change trend of hydrolase activity in different food waste was consistent.

Bacillus et al. were the dominant degraders in the food waste aerobic fermentation.